AI just broke through a major bottleneck in coin analysis

AlbumNerd

My work on AI coin analysis and grading has me closely following advancements in AI models, and the industry is evolving rapidly.

Previously, AI could only analyze static images of coins, which severely limited accuracy.

Collectors know the importance of viewing coins from multiple angles and lighting. Well...Google’s just updated Gemini 1.5 Pro to support video uploads. It can now analyze thousands of frames instead of just a few images. This is a game-changer.

I recorded a short video to check whether it would even return coherent results. I ran a quick test on a slabbed 1925 Lincoln Cent. The AI accurately identified the coin, though the grade was slightly off. That's okay, though. If I had used a high-resolution camera, set up better lighting, conducted additional tests to determine the optimal video length, and fine-tuned the model, I bet I would have gotten an accurate grade.

That's not to mention that OpenAI just released the best AI model in the market, the o1 model, which now has PhD level intelligence and reasoning. That model can't take video inputs yet, but I'm excited to test it out when it does.

In short, AI can now mimic the conditions of a human grader.

Typekat

When can we submit to AICGS?

Bikergeek

@AlbumNerd I'm not closely watching AI advancements - but I'm watching your advancements. Appreciate your work in this area - keep on keeping on!

DocBenjamin

"In short, AI can now mimic the conditions of a human grader."

Long hours, cheap booze and bloodshot eyes?

1madman

Humor me please and run Dan’s beaver coin through the AI analysis. I’m curious to know what it thinks the grade is.

https://player.vimeo.com/video/1006808292

coinkat

So the Readers’ Digest condensed version is that grading has been reduced to cheap booze…

I can’t help but feel that something was lost in the translation… but it might help set expectations

DocBenjamin

@1madman said:
Humor me please and run Dan’s beaver coin through the AI analysis. I’m curious to know what it thinks the grade is.

https://player.vimeo.com/video/1006808292

Yeah...no kidding.

CaptHenway

As the first TPG Professional Grader, I would expect to see a lot more sample grading before you say that the system works. We did 3+ months of testing before we started accepting coins.

allnewsanchor

Sounds promising. But having lived in St Louis for 7 years, I can say "I'm from Missouri. Show me."

CRHer700

AI is not an advancement that I am willing to put faith in. Many here have heard my view on the subject, so I won't go into details, but I will put forth a logical argument: If AI "can now mimic the conditions of a human grader", then if we are all complaining about how inaccurate the TPGs are, we will then complain about how inaccurate AICGS is. Not much of an improvement in my opinion.

CRHer700

Also, didn't Compugrade go under for a reason?

jmlanzaf

@CRHer700 said:
AI is not an advancement that I am willing to put faith in. Many here have heard my view on the subject, so I won't go into details, but I will put forth a logical argument: If AI "can now mimic the conditions of a human grader", then if we are all complaining about how inaccurate the TPGs are, we will then complain about how inaccurate AICGS is. Not much of an improvement in my opinion.

People are going to complain. Your logic would also argue against using any TPG or expert opinion at all.

Mr_Spud

It’s already started 😉

Crypto

AI will only consider established factors in the ether of saved data. There are too many unknown variables in the decades of unknowns and variances in strike, surfaces textures and how age and handing affects those. Not to mention the shift to market grading based off eye appeal.

I am not so sure modern grading lends its self to computer grading. thousands of beginner bag Morgan’s or Lincolns and it might make it appear closer than the art demands. I suspect they would need a modified/new standard and build the data around to work and that’s a lot of effort when it is unclear if the market would embrace it.

Jacques_Loungecoque

Can’t wait to see how CAC evaluates AI graded coins. 😉

Married2Coins

@CaptHenway said:
As the first TPG Professional Grader, I would expect to see a lot more sample grading before you say that the system works. We did 3+ months of testing before we started accepting coins.

You worked for ANACS right? That was the first authentication service so you must have been the first professional coin authenticator. ANACS was the second grading service and I believe NCI or Acugrade was either the third or forth.

Married2Coins

@AlbumNerd said:
If I had used a high-resolution camera, set up better lighting, conducted additional tests to determine the optimal video length, and fine-tuned the model, I bet I would have gotten an accurate grade.

Two little letters "IF" define computer grading for me. IMO as a long time collector the PCGS Expert was a far$e and Compugrade was a failed folly but I liked their slab. True AI grading is not going to happen. What is possible right now is a coin can be graded and fingerprinted so that it's grade will never change = computer graded coin.

Overdate

Is AI capable of evaluating all the nuances of grading? Can it accurately spot the difference between a 69 and a 70? Or a straight grade and a "details" grade? Or the four flavors of VF? Or wear vs. weak strike? Or natural vs. artificial toning? Or proof vs. DMPL? Or business strike vs. SMS vs. Satin Finish?

I expect that AI grading for some coins is feasible now, and will continue to improve over time. But I doubt that it's capable of replacing human graders just yet.

MrEureka

The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

CaptHenway

@Married2Coins said:

@CaptHenway said:
As the first TPG Professional Grader, I would expect to see a lot more sample grading before you say that the system works. We did 3+ months of testing before we started accepting coins.

You worked for ANACS right? That was the first authentication service so you must have been the first professional coin authenticator. ANACS was the second grading service and I believe NCI or Acugrade was either the third or forth.

ANACS started "Authenticating" coins in 1972, so there were several "Authenticators" there before me. I was hired as an "Authenticator' in 1978 specifically to add a "Grading" program to the existing Service, which we did on March 1, 1979.

DocBenjamin

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

Coin Finder

Fascinating!

Mjwags

I feel it could be a good tool for beginners. It may not be completely accurate but for someone with no grading experience or someone who inherits a collection, it could give them a general idea of a coins value. Something's computers just can't replace.

fathom

@Overdate said:
Is AI capable of evaluating all the nuances of grading? Can it accurately spot the difference between a 69 and a 70? Or a straight grade and a "details" grade? Or the four flavors of VF? Or wear vs. weak strike? Or natural vs. artificial toning? Or proof vs. DMPL? Or business strike vs. SMS vs. Satin Finish?

I expect that AI grading for some coins is feasible now, and will continue to improve over time. But I doubt that it's capable of replacing human graders just yet.

Yes to all those Qs you posed. It will definitely have an opinion. Some of those Qs are in dispute now with the current TPGs.

A competent evolution of AI in grading will put to shame some very expensive high grade coins in current TPG. And yes standards will evolve. AI can recode its software so to speak.

It will have its opinions on everything and yes the market may not want them.

CoffeeTime

I agree with @AlbumNerd. The advancement in AI continues at impressive speeds.

Jacques_Loungecoque

@CoffeeTime said:
I agree with @AlbumNerd. The advancement in AI continues at impressive speeds.

Impressive, yet frightening.

jmlanzaf

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

There are self driving cars on the road now. And they will continue to improve. If you used the internet in 1980, it looked nothing like it did in 1990 much less today.

MrEureka

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

None of this is going to be easy, but we're very early in the game.

Overdate

I wonder how well AI will detect counterfeits, considering that technological progress will also make counterfeits more and more difficult to detect.

MrEureka

@Overdate said:
I wonder how well AI will detect counterfeits, considering that technological progress will also make counterfeits more and more difficult to detect.

I suppose AI will do it the same way humans would, but with a larger and more accessible database for comparisons.

DocBenjamin

@jmlanzaf said:

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

There are self driving cars on the road now. And they will continue to improve. If you used the internet in 1980, it looked nothing like it did in 1990 much less today.

You are an educator. Can your position be replaced successfully by AI?

If not now, how many more years?

Cougar1978

Looking forward to the AI self grading holder.

jmlanzaf

@DocBenjamin said:

@jmlanzaf said:

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

There are self driving cars on the road now. And they will continue to improve. If you used the internet in 1980, it looked nothing like it did in 1990 much less today.

You are an educator. Can your position be replaced successfully by AI?

If not now, how many more years?

Ironically, i told my class this morning that some of my function could already be replaced. I used ChatGPT to ask specific content questions that we had discussed this week. The result was a well structured description of what we had talked about in class. Now, my replaceability goes down as the complexity of the material goes up... for now.

What really changes, ultimately, is the necessary skills for a student. For example, there is little need to either teach or test on facts that your cell phone can provide. There is, however, a need to properly query the AI to get good responses as well as critical evaluation skills required to assess the response.

I beg my students to use AI for homework and learning. I insist they use it to help write papers. I implore them to consider how their career will be defined by their ability to use AI.

BStrauss3

@CRHer700 said:
Also, didn't Compugrade go under for a reason?

They couldn't make it work and most coins were hand graded...

It's also a very different product. Compugrade made models of the coins and divided up the field and design into dozens or 100s of zones and measured "something" in each zone that was manipulated to create the grade.

https://patents.google.com/patent/US4899392A/en

There is also our hosts' patent: https://patents.google.com/patent/US5220614A/en

leothelyon

If there is one area of coin grading the cgc's have failed to master for over 25+ years is determining/calculating the strength/completion of the strike/details implementing that criteria, a must/important factor to receive a MS66 and higher grade. Yeah, sure, if AI coin grading can do that, determine the depth of the details on a coin.....although....I just love landing a 4, 5 even a 6 quality calibur coin for $50 anytime of the week. I've been doing so for all those last 34+ years so I wouldn't really welcome a change to that. And some of you have already seen my latest with frost and lathe lines, etc

Leo

MrEureka

@BStrauss3 said:

@CRHer700 said:
Also, didn't Compugrade go under for a reason?

They couldn't make it work and most coins were hand graded...

It's also a very different product. Compugrade made models of the coins and divided up the field and design into dozens or 100s of zones and measured "something" in each zone that was manipulated to create the grade.

https://patents.google.com/patent/US4899392A/en

There is also our hosts' patent: https://patents.google.com/patent/US5220614A/en

"Feb 6, 1990"

jacrispies

@Jacques_Loungecoque said:
Can’t wait to see how CAC evaluates AI graded coins. 😉

Introducing the new RED CAC sticker!!! Now symbolizing grossly overgraded coins

CRHer700

@jmlanzaf said:

@CRHer700 said:
AI is not an advancement that I am willing to put faith in. Many here have heard my view on the subject, so I won't go into details, but I will put forth a logical argument: If AI "can now mimic the conditions of a human grader", then if we are all complaining about how inaccurate the TPGs are, we will then complain about how inaccurate AICGS is. Not much of an improvement in my opinion.

People are going to complain. Your logic would also argue against using any TPG or expert opinion at all.

I don't use any TPG at the current moment. Expert opinion is just that, opinion. Personally, I would rather have a human's opinion than a bot's.

PeakRarities

Readers will look back at this thread 20 years ago and laugh. It’s not a matter of “if”, but “when”. If the goal is only to match the “accuracyne” of human graders, ai technology's is already more than capable of doing so.

The issue is that when the technology is applied with the proper data sets, it would expose the current system for what it is. To be abundantly clear, I use the word “expose” not because it’s some big conspiracy, just that the system is inherently flawed, as most of us know. The first respectable entity to implement an AI system will state, authoritatively, that “Our system is the arbiter of truth, and thus, only our grade is the coins true and everlasting grade.” 🤴🏼

This presents a bit of a conundrum for PCGS, with respect to optics. They would effectively have to invalidate the system that served as the foundational bedrock, upon which their entire business model was built. They’ll have to be thoughtful and delicate in their approach.

CaptHenway

As an example of the limitations of modern technology, I present to you the McDonald’s milkshake.

blitzdude

@Jacques_Loungecoque said:
Can’t wait to see how CAC evaluates AI graded coins. 😉

Probably with a special AI approved sticker on top of their silly sticker. Then the cacs cult can pay 2x what they pay now. Crazy world. RGDS!

PeakRarities

I must say it’s quite amusing to watch someone fight an imaginary battle in their head.

DocBenjamin

@PeakRarities said:
Readers will look back at this thread 20 years ago and laugh. It’s not a matter of “if”, but “when”. If the goal is only to match the “accuracyne” of human graders, ai technology's is already more than capable of doing so.

The issue is that when the technology is applied with the proper data sets, it would expose the current system for what it is. To be abundantly clear, I use the word “expose” not because it’s some big conspiracy, just that the system is inherently flawed, as most of us know. The first respectable entity to implement an AI system will state, authoritatively, that “Our system is the arbiter of truth, and thus, only our grade is the coins true and everlasting grade.” 🤴🏼

This presents a bit of a conundrum for PCGS, with respect to optics. They would effectively have to invalidate the system that served as the foundational bedrock, upon which their entire business model was built. They’ll have to be thoughtful and delicate in their approach.

Twenty years from now, we may be trading in stones.

Married2Coins

@Overdate said:
Is AI capable of evaluating all the nuances of grading? Can it accurately spot the difference between a 69 and a 70? Or a straight grade and a "details" grade? Or the four flavors of VF? Or wear vs. weak strike? Or natural vs. artificial toning? Or proof vs. DMPL? Or business strike vs. SMS vs. Satin Finish?

I expect that AI grading for some coins is feasible now, and will continue to improve over time. But I doubt that it's capable of replacing human graders just yet.

IMO, the line between 69 &70 is the easiest line in the world for a computer to define because any defect (either a tiny nick, mint made strike through, or spot) would kill the perfect grade. Unfortunately, that would be too critical compared to today's commercial "eye-ball grading" and dealers would hate it.

A friend sent a MS-70 1986 SE in for review with a nick on the letters (a $40 coin when graded long ago and now worth much more) and was told that was the TPGS standard for 70! I'd bet this has happened to others here.

Married2Coins

@CaptHenway said:

@Married2Coins said:

@CaptHenway said:
As the first TPG Professional Grader, I would expect to see a lot more sample grading before you say that the system works. We did 3+ months of testing before we started accepting coins.

You worked for ANACS right? That was the first authentication service so you must have been the first professional coin authenticator. ANACS was the second grading service and I believe NCI or Acugrade was either the third or forth.

ANACS started "Authenticating" coins in 1972, so there were several "Authenticators" there before me. I was hired as an "Authenticator' in 1978 specifically to add a "Grading" program to the existing Service, which we did on March 1, 1979.

That makes ANACS the First Authentication Service and the second Third-Party Grading Service since the second Authentication Service (INSAB) began grading coins for its customers several months before ANACS. One of the original graders of that company posts on another forum where I post.

Married2Coins

@jmlanzaf said:

@DocBenjamin said:

@jmlanzaf said:

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

There are self driving cars on the road now. And they will continue to improve. If you used the internet in 1980, it looked nothing like it did in 1990 much less today.

You are an educator. Can your position be replaced successfully by AI?

If not now, how many more years?

Ironically, i told my class this morning that some of my function could already be replaced. I used ChatGPT to ask specific content questions that we had discussed this week. The result was a well structured description of what we had talked about in class. Now, my replaceability goes down as the complexity of the material goes up... for now.

What really changes, ultimately, is the necessary skills for a student. For example, there is little need to either teach or test on facts that your cell phone can provide. There is, however, a need to properly query the AI to get good responses as well as critical evaluation skills required to assess the response.

I beg my students to use AI for homework and learning. I insist they use it to help write papers. I implore them to consider how their career will be defined by their ability to use AI.

But AI controls what is learned. We can't post here using examples of the way Google and the press block or change actual facts in politics, Covid, etc. Thus, it can change grading in any way it wants. And who needs AI computer grading it there will be an evolving lack of standardization? Give me PCGS!

CaptHenway

We’ll just skip this part.

percyb

Thanks much for your AI insights on grading.
Can it identify counterfeit
coins? Curious in Chicagoland.

Married2Coins

They take pictures of coins at the TPGS and that must help catch fakes. AI could do the same.

jmlanzaf

@Married2Coins said:

@jmlanzaf said:

@DocBenjamin said:

@jmlanzaf said:

@DocBenjamin said:

@MrEureka said:
The most intriguing part about AI grading is that we will be told up front that the standards will be constantly evolving. Scary in a way, but also liberating in a way.

it won't work though. 100 MS 65 Morgan Dollars all have different elements of deviation from perfection.

Similarly, that is why the self driving car is virtually bust. Too many factors to process.

There are self driving cars on the road now. And they will continue to improve. If you used the internet in 1980, it looked nothing like it did in 1990 much less today.

You are an educator. Can your position be replaced successfully by AI?

If not now, how many more years?

Ironically, i told my class this morning that some of my function could already be replaced. I used ChatGPT to ask specific content questions that we had discussed this week. The result was a well structured description of what we had talked about in class. Now, my replaceability goes down as the complexity of the material goes up... for now.

What really changes, ultimately, is the necessary skills for a student. For example, there is little need to either teach or test on facts that your cell phone can provide. There is, however, a need to properly query the AI to get good responses as well as critical evaluation skills required to assess the response.

I beg my students to use AI for homework and learning. I insist they use it to help write papers. I implore them to consider how their career will be defined by their ability to use AI.

But AI controls what is learned. We can't post here using examples of the way Google and the press block or change actual facts in politics, Covid, etc. Thus, it can change grading in any way it wants. And who needs AI computer grading it there will be an evolving lack of standardization? Give me PCGS!

That actually doesn't change anything that i said. Try to not use the internet and see how you perform relative to your peers.

The application developer also has the ability to use the AI as an API and therefore has control over how the LLM is used and what information it receives. So, for example, PCGS AI would largely be in the hands of PCGS not open AI. If standards charged, it would be because the TPGS wanted the change.

Personally, i don't think AI grading will be any better than human grading. It might be cheaper and faster. It might be equal in precision. But it has the same basic problems that limits human grading, primarily the uniqueness of each coin.

Onastone

“Open the pod bay doors, HAL”